Notebooks

Notebooks (⌘7) is a Jupyter environment inside the app, running on the container’s own Python interpreter (SimNIBS + TI-Toolbox), so nothing needs installing.

Notebooks are ordinary .ipynb files in your project, under <project>/code/ti-toolbox/notebooks/. Saving round-trips the document, so a notebook edited here and a notebook edited in Jupyter Lab produce the same file.

The seeded worked example, run against a real project. Each cell makes the same call as one page of the app.

Example workflow

Example workflow on the freely available subject ernie.

1. Load

Point at a project folder; fetch_ernie downloads the example subject into it.

from tit import get_path_manager
from tit.examples import fetch_ernie
from tit.pre import discover_subjects

PROJECT = "/path/to/your/project"
SUBJECT = "ernie"

pm = get_path_manager(PROJECT)
fetch_ernie(PROJECT)
print(discover_subjects(PROJECT))

2. Pre-process

Build the head model with charm. Ernie ships with one, so this is skipped there.

from tit.pre import run_pipeline, check_m2m_exists

if not check_m2m_exists(PROJECT, SUBJECT):
    run_pipeline([SUBJECT], convert_dicom=True, create_m2m=True)

3. Optimise

Flex-search for electrode positions that drive a 10 mm sphere at an MNI coordinate.

from tit.opt import FlexConfig, run_flex_search

flex = FlexConfig(
    subject_id=SUBJECT,
    goal="mean",
    postproc="max_TI",
    current_mA=1.0,
    electrode=FlexConfig.ElectrodeConfig(shape="ellipse", dimensions=[8.0, 8.0]),
    roi=FlexConfig.SphericalROI(x=-37.0, y=-21.0, z=58.0, radius=10.0, use_mni=True),
)
result = run_flex_search(flex)
print(result.best_value, result.output_folder)

4. Simulate

Save a montage by name, then simulate it.

from tit.sim import upsert_montage, load_montages, SimulationConfig, run_simulation

NET = "EEG10-10_UI_Jurak_2007.csv"
MONTAGE = "L_Insula"
upsert_montage(eeg_net=NET, montage_name=MONTAGE, mode="U",
               electrode_pairs=[["F7", "TP7"], ["FT7", "P7"]])

sim = SimulationConfig(
    subject_id=SUBJECT,
    montages=load_montages(montage_names=[MONTAGE], eeg_net=NET),
    intensities=[1.0, 1.0],
)
run_simulation(sim, overwrite=True)

5. Analyse

Read the field in a cortical region.

from tit.analyzer import Analyzer

analyzer = Analyzer(subject_id=SUBJECT, simulation=MONTAGE, space="mesh")
roi = analyzer.analyze_cortex(atlas="DK40", region="lh.insula")
print(roi.roi_mean, roi.roi_max, roi.roi_focality)

Everything else is in the API reference. A runnable copy of this workflow is the example notebook below.

The example notebook

Every project gets one seeded notebook the first time you open the page, examples/example_workflow.ipynb. It walks the app page by page — project, pre-processing, Optimizer, Simulator, Analyzer, results — one cell per page, showing the single tit API call each page’s Run button makes. Cell 1 fetches the ernie example subject (Example Data), so it runs without your own data. The same calls are walked through in The workflow in code above.

⬇ Download example_workflow.ipynb    View on GitHub

When a TI-Toolbox update ships a newer version, your copy is refreshed only if you never edited it; an edited copy is yours. Deleting it is permanent; there is no “restore example”.